Acceleration of sensitivity analysis by distributed computation

Area of Science:

• Computational biology
• Systems biology
• Cardiac electrophysiology

Background:

• Developing intricate mathematical models of cardiac myocytes necessitates substantial computational resources.
• Sensitivity analyses are crucial for understanding and validating complex biological models.
• Advancements in distributed computing offer potential solutions for high-performance biological simulations.

Purpose of the Study:

• To enhance the computational efficiency of cardiac myocyte modeling.
• To accelerate sensitivity analyses for complex biological models.
• To develop a practical, accessible system for distributed biological simulations.

Main Methods:

• Utilized the simBio biological simulator for mathematical modeling of cardiac myocytes.
• Implemented a distributed computing approach using a cluster of 11 personal computers (PCs).
• Integrated the Jini open-source package for network orchestration of the PC cluster with simBio.

Main Results:

• Achieved a 13-fold reduction in computation time for 504 models compared to a single PC.
• Demonstrated a linear decrease in computation time with the addition of PCs, confirming scalability.
• The integrated system provides researchers with instant feedback through visualization, accelerating model-driven studies.

Conclusions:

• Distributed computing using a PC cluster significantly accelerates sensitivity analyses for complex biological models.
• The developed simBio-Jini system offers an efficient and accessible solution for high-performance biological simulations.
• This approach facilitates faster model development, validation, and utilization in cardiac research.